This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-m-v2.0 on the train dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: Snowflake/snowflake-arctic-embed-m-v2.0
Maximum Sequence Length: 8192 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- train

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 8192, 'do_lower_case': False, 'architecture': 'GteModel'})
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("BjarneNPO/finetune_21_08_2025_11_36_21")
# Run inference
queries = [
    "fragt wie der Stand zu dem aktuellen Problem ist",
]
documents = [
    'In Klärung mit der Kollegin - Das Problem liegt leider an deren Betreiber. Die sind aber informiert und arbeiten bereits daran',
    'findet diese in der Übersicht der Gruppen.',
    'Userin muss sich an die Bistums IT wenden.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 768] [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 0.2882, -0.0027, -0.0251]])

Evaluation

Metrics

Information Retrieval

Dataset: Snowflake/snowflake-arctic-embed-m-v2.0
Evaluated with scripts.InformationRetrievalEvaluatorCustom.InformationRetrievalEvaluatorCustom with these parameters:
json
1{ 2 "query_prompt_name": "query", 3 "corpus_prompt_name": "query" 4}

Metric	Value
cosine_accuracy@1	0.1825
cosine_accuracy@3	0.4015
cosine_accuracy@5	0.5036
cosine_accuracy@10	0.708
cosine_precision@1	0.1825
cosine_precision@3	0.2482
cosine_precision@5	0.2394
cosine_precision@10	0.2022
cosine_recall@1	0.0133
cosine_recall@3	0.0277
cosine_recall@5	0.0478
cosine_recall@10	0.0836
cosine_ndcg@10	0.2223
cosine_mrr@10	0.3408
cosine_map@100	0.0965

Training Details

Training Dataset

train

Dataset: train
Size: 19,964 training samples
Columns: query and answer
Approximate statistics based on the first 1000 samples:
query answer
type string string
details
min: 4 tokens
mean: 27.77 tokens
max: 615 tokens
min: 3 tokens
mean: 22.87 tokens
max: 151 tokens

Samples:

query	answer
`Wie kann man die Jahresurlaubsübersicht exportieren?`	`über das 3 Punkte Menü rechts oben. Mitarbeiter auswählen und exportieren`
`1. Vertragsabschlüsse werden nicht übertragen`
2. Kinder kommen nicht von nach
3. Absage kann bei Portalstatus nicht erstellt werden.	`Ticket`
Userin gebeten sich an den Support zu wenden, da der Fehler liegt.
`Wird im Anmeldeportal nicht gefunden.`	`Die Schnittstelle war noch nicht aktiviert und Profil ebenfalls nicht.`

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
gradient_accumulation_steps: 8
learning_rate: 2e-05
num_train_epochs: 10
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: True
tf32: False
load_best_model_at_end: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 8
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 10
max_steps: -1
lr_scheduler_type: cosine
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: False
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: True
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_tokens_per_second: False
include_num_input_tokens_seen: False
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Snowflake/snowflake-arctic-embed-m-v2.0_cosine_ndcg@10
0.0641	10	2.1581	-
0.1282	20	2.259	-
0.1923	30	2.0893	-
0.2564	40	1.9604	-
0.3205	50	1.9998	-
0.3846	60	1.8531	-
0.4487	70	1.7507	-
0.5128	80	1.7337	-
0.5769	90	1.6102	-
0.6410	100	1.5674	-
0.7051	110	1.4687	-
0.7692	120	1.4123	-
0.8333	130	1.4028	-
0.8974	140	1.2895	-
0.9615	150	1.2337	-
1.0	156	-	0.2480
1.0256	160	1.1688	-
1.0897	170	1.1417	-
1.1538	180	1.2043	-
1.2179	190	1.0708	-
1.2821	200	1.0728	-
1.3462	210	1.1041	-
1.4103	220	1.0581	-
1.4744	230	0.9842	-
1.5385	240	0.9806	-
1.6026	250	0.9709	-
1.6667	260	0.9564	-
1.7308	270	0.9005	-
1.7949	280	0.908	-
1.8590	290	0.8904	-
1.9231	300	0.8536	-
1.9872	310	0.8062	-
2.0	312	-	0.2458
2.0513	320	0.7144	-
2.1154	330	0.7484	-
2.1795	340	0.8055	-
2.2436	350	0.7487	-
2.3077	360	0.8139	-
2.3718	370	0.7268	-
2.4359	380	0.7472	-
2.5	390	0.7173	-
2.5641	400	0.7024	-
2.6282	410	0.6622	-
2.6923	420	0.6446	-
2.7564	430	0.6699	-
2.8205	440	0.6792	-
2.8846	450	0.5952	-
2.9487	460	0.5851	-
3.0	468	-	0.2282
3.0128	470	0.5691	-
3.0769	480	0.5537	-
3.1410	490	0.583	-
3.2051	500	0.5815	-
3.2692	510	0.5497	-
3.3333	520	0.5772	-
3.3974	530	0.5536	-
3.4615	540	0.5584	-
3.5256	550	0.5303	-
3.5897	560	0.4791	-
3.6538	570	0.4992	-
3.7179	580	0.4669	-
3.7821	590	0.5269	-
3.8462	600	0.4864	-
3.9103	610	0.4765	-
3.9744	620	0.4192	-
4.0	624	-	0.2232
4.0385	630	0.3851	-
4.1026	640	0.4163	-
4.1667	650	0.4397	-
4.2308	660	0.4371	-
4.2949	670	0.4457	-
4.3590	680	0.4207	-
4.4231	690	0.4137	-
4.4872	700	0.4005	-
4.5513	710	0.4093	-
4.6154	720	0.3574	-
4.6795	730	0.3696	-
4.7436	740	0.3806	-
4.8077	750	0.4206	-
4.8718	760	0.3342	-
4.9359	770	0.3503	-
5.0	780	0.3373	0.2308
5.0641	790	0.3064	-
5.1282	800	0.3173	-
5.1923	810	0.3537	-
5.2564	820	0.3226	-
5.3205	830	0.3398	-
5.3846	840	0.3268	-
5.4487	850	0.3581	-
5.5128	860	0.3097	-
5.5769	870	0.3019	-
5.6410	880	0.2782	-
5.7051	890	0.2907	-
5.7692	900	0.304	-
5.8333	910	0.3223	-
5.8974	920	0.2812	-
5.9615	930	0.2776	-
6.0	936	-	0.2341
6.0256	940	0.2451	-
6.0897	950	0.2607	-
6.1538	960	0.2769	-
6.2179	970	0.2691	-
6.2821	980	0.269	-
6.3462	990	0.2767	-
6.4103	1000	0.2663	-
6.4744	1010	0.2716	-
6.5385	1020	0.2531	-
6.6026	1030	0.2387	-
6.6667	1040	0.2565	-
6.7308	1050	0.2385	-
6.7949	1060	0.2799	-
6.8590	1070	0.2442	-
6.9231	1080	0.2448	-
6.9872	1090	0.2307	-
7.0	1092	-	0.2453
7.0513	1100	0.2001	-
7.1154	1110	0.2114	-
7.1795	1120	0.2407	-
7.2436	1130	0.2436	-
7.3077	1140	0.2405	-
7.3718	1150	0.2396	-
7.4359	1160	0.2442	-
7.5	1170	0.2334	-
7.5641	1180	0.2355	-
7.6282	1190	0.2007	-
7.6923	1200	0.203	-
7.7564	1210	0.237	-
7.8205	1220	0.241	-
7.8846	1230	0.2149	-
7.9487	1240	0.2119	-
8.0	1248	-	0.2217
8.0128	1250	0.2051	-
8.0769	1260	0.1919	-
8.1410	1270	0.1911	-
8.2051	1280	0.2334	-
8.2692	1290	0.2101	-
8.3333	1300	0.2211	-
8.3974	1310	0.2028	-
8.4615	1320	0.2314	-
8.5256	1330	0.1962	-
8.5897	1340	0.181	-
8.6538	1350	0.2027	-
8.7179	1360	0.1853	-
8.7821	1370	0.2206	-
8.8462	1380	0.213	-
8.9103	1390	0.2239	-
8.9744	1400	0.1901	-
9.0	1404	-	0.2217
9.0385	1410	0.1676	-
9.1026	1420	0.1936	-
9.1667	1430	0.2123	-
9.2308	1440	0.2025	-
9.2949	1450	0.2062	-
9.3590	1460	0.2191	-
9.4231	1470	0.2037	-
9.4872	1480	0.2053	-
9.5513	1490	0.2124	-
9.6154	1500	0.1852	-
9.6795	1510	0.198	-
9.7436	1520	0.1924	-
9.8077	1530	0.2375	-
9.8718	1540	0.1906	-
9.9359	1550	0.2056	-
10.0	1560	0.1826	0.2223

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.11
Sentence Transformers: 5.1.0
Transformers: 4.55.2
PyTorch: 2.8.0+cu129
Accelerate: 1.10.0
Datasets: 3.6.0
Tokenizers: 0.21.4

Citation

BibTeX

Sentence Transformers

bibtex

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

bibtex

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

	query	answer
type	string	string
details	min: 4 tokens mean: 27.77 tokens max: 615 tokens	min: 3 tokens mean: 22.87 tokens max: 151 tokens